Some Interesting Facts About Sandy bridge

Joined
Aug 26, 2010
Messages
23
Hi guys, I found an interesting article about Sandy Bridge.. Just want to share it to all of you..:)

We spent some time fiddling around with a couple of Sandy Bridge processors and several motherboards based on different chipsets, during which we found some crucial facts that could prove to be vital information for would-be-Sandy Bridge users (in particular, those who are tempted by the new features and the seemingly amazing degree of overclocklability offered by these processors). Here are some of the most important:

1. The new Sandy Bridge CPUs (in LGA 1155 packaging) are NOT backward compatible with LGA 1156 sockets
soket.jpg


This one is simple and obvious, of course. If you want to use a Sandy Bridge, you’ll have to plug it on one of the new LGA 1155 motherboards. These processors are not compatible with older, LGA 1156-based motherboards. The only thing that remains similar between LGA 1156 and LGA 1155 platforms are their HSF cooler mountings. At least you get to install your trusty LGA 1156 cooler on the new motherboard, but don’t expect anything more than that.

2. K-Series and Non-K-Series are Not Created Equal
No, it’s not like heaven and earth, really. The only difference between “Ks” and “Non-Ks” is the unlocked multiplier on the “K series” Sandy Bridges. Everything else, from clockspeed right down to each hardware feature, remains the same. As you have probably known, an unlocked multiplier allows for a far greater flexibility in overclocking attempts since the core frequency can be increased without affecting other parameters.

The “K” designation as in the “Core i5-2500K”, therefore, serves to differentiate between processors with and without the unlocked multiplier. In the case of the Core i5-2500K, for example, it has a twin sibling that ships with a locked multiplier. This processor is simply called the Core i5-2500.

3. Sandy Bridge’s Bus Speed Overclocking is SEVERELY Limited
This is the basic overclocking formula: CPU Clock = Bus Speed x CPU Multiplier. The bus speed, or commonly referred to as the “BCLK” or “Base Clock”, of previous Intel processors (Lynnfield, Bloomfield, Gulftown, and also Clarkdale) could be adjusted quite flexibly between 133 MHz to 220 MHz or even 250 MHz using extreme cooling. With a BCLK adjustment range of around 50%, you don’t need an unlocked multiplier to reach unbelievably high frequencies. This is not the case with Sandy Bridge processors, as their BCLK range is VERY limited.

Sandy Bridges have a default BCLK frequency of 100Mhz, but one can only expect a 4-5% adjustment to that value. In average, our Sandy Bridge samples hit a BCLK limitation at 104-105 MHz. Even by resorting to extreme cooling methods, we could only get 110 MHz at maximum. It means that you will have to rely on multiplier adjustment to overclock a Sandy Bridge processor. Also, since unlocked multiplier can only be found in “K-series” Sandy Bridges, either you like it or not, you will have to buy one of those to get a significant overclocking margin. You can still apply some degree of overclocking on the “Non-K series” processors, but you will be limited to just 400 MHz above the maximum turbo frequency.Such limitation prevents users from overclocking a lower-end processor to match the performance of its higher-end variants, like what happens when you overclock a Core i7 920 to the same speed as Core i7 975 Extreme’s. That way, Intel keeps the sales of its high-end CPUs from being cannibalized by its own value series processors. By the time we finished writing this article, Intel has yet to confirm whether the BCLK limitation is “inadvertently” caused by the new architecture or if it is deliberately implemented. Either way, the Sandy Bridge is still limited when it comes to BCLK overclocking, and that could be disappointing to some types of users.

4. There are Actually Significant Differences Between H67 and P67-based Motherboards
Surely you still remember how the H55 chipset is not so different compared to its sibling, the P55. The only thing that distinguishes them is H55’s ability to take advantage of on-CPU graphics chip in Clarkdale series processors (the Core i3, i5 5xx, and i5 6xx). This time, Intel saw it fitting to incorporate some major changes to set the H67 and P67 further apart. Take a look at the table below:

P67vsH67.jpg


For some reason, the H67 seems to be limited in certain aspects. It appears that Intel doesn’t want H67 users to overclock their CPUs, either through multiplier adjustment or BCLK manipulation. They also prevent the DRAM speed on H67 from going beyond 1333 MHz. You can still overclock the IGP (Integrated Graphics Processor) though, but not the main CPU cores.

So, for those of you who are drawn to “K” Sandy Bridges by its multiplier overclocking potential, (also, owners of high-frequency RAM modules) you’re better off with a P67 –based motherboard, at least for now. Remember that you cannot overclock the CPU on a H67 motherboard. However, it’s worth remembering that Intel did not include on-CPU graphics support when they designed the P67, so that part of your Sandy Bridge processor will pretty much be wasted should you decide to pair the processor with a motherboard based on this chipset. If you want to utilize Sandy Bridge’s IGP, you’ll have to resort to a H67 with all its overclocking restrictions. Pretty confusing isn’t it?

5. Sandy Bridge Chipsets Have Limited DRAM Speed Support
Have you had enough with the “limitation” word already? We hope you haven’t, for we have one more to tell you about. As you have probably guessed, Sandy Bridge’s tiny BCLK adjustment range severely limits the frequency selection available for your RAM modules. H67 platforms are the worst in this case, supporting a maximum RAM speed only 1333 MHz. Fortunately, Intel was kind enough to provide P67 users with a bit more freedom by allowing them to select between more varied RAM frequency settings: 1333Mhz, 1600Mhz, 1866Mhz, and 2133Mhz. You will be able to run your memory at up to 2133 MHz on a P67 motherboard, but that’s it. If your RAM modules are rated at speeds other than those mentioned above, such as DDR3-1800, DDR3-2200, DDR3-2400, etc., you will likely have a hard time trying to adjust the BCLK to allow your RAM to run at its highest speed. Sandy Bridges use this formula to determine memory frequency:

DRAM Speed = BCLK * DRAM Multiplier, which means that the range of Sandy Bridge’s RAM frequency setting is restricted by its limited BCLK adjustment.

Take this, for example:

To run his RAM at DDR3-2400, a user would have to set the DRAM speed at 2133 MHz, and then manually raising the BCLK to strong>112.5 Mhz. Since Sandy Bridges hit a BCLK limitation at 104-105 MHz on average, that will be very difficult –if not impossible- to achieve.

It also means that the Sandy Bridge has different XMP (Xtreme Memory Profile) settings. On a Sandy Bridge platform, XMP modules will step down to run at the closest rated frequency (under its own rated speed) as specified by Intel: at 1333, 1600, 1866, or 2133Mhz). Suppose you have a DDR3-2200 MHz rating in your RAM’s XMP profile, you can only expect it to run at DDR3-2133 MHz on a Sandy Bridge motherboard. DDR3-1800 MHz XMP Modules will be similarly stepped down to DDR3-1600, as will the other modules with different rated speeds.

Source: http://www.jagatreview.com/2011/01/the-top-five-things-you-should-know-about-sandybridge/

Additional information:
Please add to the list of important info:

LGA 1155 is Intel's new midrange socket. Midrange socket platforms come with limitations. If you attempt to use midrange equipment at extreme values, expect disappointment and failures. If you want a high-end socket platform with not so many limitations, wait for LGA 2011.
 
Last edited:
No really novel information there, but it's probably useful to see it all in one place for people new to the whole SB thing.
 
SB is more restricted than nehalem, but at the same time straightforward.

So if you want the most out of SB as of right now. You're going to get a 2600k + a P67 mobo + 2133 ram to start with
 
Thanks Lord_of_the_Obvious, some more facts:

#1 - The max safe voltage is 1.52v (as stated in Intel's datasheet, page 78) Wait for an update on this one!

#2 - The Tcase is lower than previous chips @ 72.6c (this is where thermal throttling starts kicking in, the CPU version of a siesta).​


Stay <= 1# & #2 and you should have allot of fun!

Keep the facts rolling..
 
Last edited:
This looks like a good thread to drop this in

http://forums.overclockers.co.uk/showpost.php?p=18190523&postcount=103

Hi there


Right guys myself and our technical guys have spent the entire weekend and this morning in discussions with Intel regarding the alarming amount of reports of Sandybridge CPU's dying and have been conducting our own testing as have Intel to find out what is a definite no no.


Sandybridge maximum safe voltages

Core Voltage - Not recommended too exceed 1.38v, doing so could kill the CPU, we therefor recommend a range of 1.325-1.350v if overclocking.
Memory Voltage - Intel recommend 1.50v plus/minus 5% which means upto 1.58v is the safe recommended limit. In our testing we have found 1.65v has caused no issues.
BCLK Base Clock - This is strictly a NO, anyone using base clock overclocking could/will cause damange to CPU/Mainboard. (Set manually to 100)
PLL Voltage - Do not exceed 1.9v!!



Processor - Basically we recommend customers not to exceed 1.35v to play it safe, all our bundles are set at 1.3250v or lower, any competitors offering bundles above 4.6GHz you should be enquiring as to what voltage they are using as we believe anything over 1.38v will limit CPU lifespan and anything over 1.42v will likely kill the CPU or severely limit its lifespan.

Memory - Intel recommend 1.50v plus/minus 5% which means 1.60v is the ideal safe maximum, but we have found in our testing all 1.65v memory is fine. We have also found most new 1.65v like Corsair XMS3 will run at its rated timings with just 1.50-1.55v which is well within Intel specifications. So people upgrading to Sandybridge you can still use your old DDR3, but we do recommend you run it at 1.60v or less. We are shipping most of our bundles which feature Corsair XMS at 1.50v-1.55v at rated timings. We've also discussed with Asus and MSI regarding voltages for memory and they also confirm in their testing 1.65v caused no issues with reliability.

Base Clock - To put it simple if you value the life of your components, do not overclock using base clock!

PLL Voltage - Again do not exceed 1.9v!


These are just guidelines we recommend you follow, if you want to push more voltage through your CPU's then just be aware they could die on you. Your warranty is un-affected and we will honor any CPU's that die, we just won't ask questions as to how you killed them.

Not all CPU's are as fragile as others, we have experimented upto 1.50v Vcore and 1.70v memory and had zero issues with reliability, so it seems some of fine when pushing hard.

Spread the word.
 
Very interesting. If it is true it is a bit of a cock up on Intel's behalf, hopefully they should release a press release quickly to clear this up since all these dead chips will start eating into their margins.
 
Very interesting. If it is true it is a bit of a cock up on Intel's behalf, hopefully they should release a press release quickly to clear this up since all these dead chips will start eating into their margins.
LOL

Enthusiasts who are extreme overclockers are a tiny, tiny percentage of Intel's sales. The vast bulk of SB chips will go into OEM systems. And most of the regular overclockers will be scared when the voltage range turns the color red in BIOS as it's increased passed maximum. The fact is that most people don't overclock CPUs at all.

Anyways, Intel has been conservative with voltages in the last generation and later increased the range in the documentation. It might be an artificially low limit again. Or not.
 
I was under the impression over clocking was becoming more main stream, you don't even have to leave the safety of windows now to overclock, loads of overclocking turbo modes.

Yes 1.52v did strike me as rather high at first (I never pushed my 65mn q6600 past 1.5v, which is 4x the size of these babies). I can't wait to hear what Intel has to say about this.

PS: VT-d really is a server only function (same as they don't support ECC memory), these will still handle visualisation just fine.
 
Last edited:
LOL

Enthusiasts who are extreme overclockers are a tiny, tiny percentage of Intel's sales. The vast bulk of SB chips will go into OEM systems. And most of the regular overclockers will be scared when the voltage range turns the color red in BIOS as it's increased passed maximum. The fact is that most people don't overclock CPUs at all.

Anyways, Intel has been conservative with voltages in the last generation and later increased the range in the documentation. It might be an artificially low limit again. Or not.

It seems that way with some of the OC results I've been seeing fly around XS the past couple of days. People have been pushing 1.5v on Air.
 
LOL

Enthusiasts who are extreme overclockers are a tiny, tiny percentage of Intel's sales.

And yet we have big mouths and have the ears of many non-technical customers. You'd think Intel would have figured that out by now.
 
And yet we have big mouths and have the ears of many non-technical customers. You'd think Intel would have figured that out by now.

It's been my experience that normal people would still rather listen to the the Best Buy employee over us.
 
1.35V seems awfully low for a max voltage - looking at the results floating around almost everyone is using something in the 1.4 range. I noticed that whoever wrote that threw in a plug for how their systems use a lower voltage - makes me question their motives a little bit.
 
I haven't bothered to look at the SB data sheets yet, but 1.35v is probably what it states. Don't shoot the messenger for telling the truth. ;)
 
I haven't bothered to look at the SB data sheets yet, but 1.35v is probably what it states. Don't shoot the messenger for telling the truth. ;)

See post #5 of this thread :rolleyes:

thanks for this. even though a forum post is not the most reliable source of information, i'll play it safe and assume there's some merit to it.

True, but the guy who posted it is an employee from a shop that has been selling i7-2's before launch (since Wednesday). Also, that happens to be the same place I brought mine from :(, yet they haven't sent me an email warning me of this. I think the best bet is to tread carefully until Intel speaks about this so-called issue.
 
1.35v is too low. 1.6-1.7 will work but is going to compromise the life expectancy a bit, 1.8 is a definitive NO. I've taken my 2600K as high as 1.55 in the BIOS for extended periods so you're not going to kill it unless you intentionally try to.
I have my 2600K folding@home at 4.7GHz with a BIOS indicated 1.45v, or about 1.4 after vdroop under load. It's been running this for over a week now with no issue, only about 20 minutes every two days does it get a break from full load.

Going over 1.35v heat starts to increase very quickly. By 1.45v you're probably at or very close to the CPU's multi wall anyway.
 
1.35v is too low. 1.6-1.7 will work but is going to compromise the life expectancy a bit, 1.8 is a definitive NO. I've taken my 2600K as high as 1.55 in the BIOS for extended periods so you're not going to kill it unless you intentionally try to.


So because you have had yours at 1.55V for a few days means that everyone else can and should have no problem running it forever. Wow! Intel should fire Intel and replace it with YOU!


:rolleyes:
 
1.35v is too low. 1.6-1.7 will work but is going to compromise the life expectancy a bit,
Not a bit, 1.7v will kill the chip in a pretty short period of time. That's called a suicide run.
 
So because you have had yours at 1.55V for a few days means that everyone else can and should have no problem running it forever. Wow! Intel should fire Intel and replace it with YOU!

:rolleyes:

Yes, that's exactly what I said. Thanks for clarifying that for everyone. :rolleyes:

My information comes directly from Intel. You never know who you're trying to insult on the internet.

Thanks.
 
Yes, that's exactly what I said. Thanks for clarifying that for everyone. :rolleyes:

My information comes directly from Intel. You never know who you're trying to insult on the internet.

So you work for Intel? Or you just read their white paper?
 
So you work for Intel? Or you just read their white paper?

white paper says 1.52V, so that leads me to believe he has some other source of 'information'. regardless, my chip isn't breaking 1.35V until more time has passed and it's been proven safe.
 
Please add to the list of important info:

LGA 1155 is Intel's new midrange socket. Midrange socket platforms come with limitations. If you attempt to use midrange equipment at extreme values, expect disappointment and failures. If you want a high-end socket platform with not so many limitations, wait for LGA 2011.
 
http://www.xtremesystems.org/forums/showpost.php?p=4699227&postcount=156
Voltages:
Now let&#8217;s move to voltages, this is very important and I think anyone looking to overclock should have a thorough introduction to the variety of voltages that the user is able to control.
Make sure to pay attention to my tid-bits on Vcore(VCC) and VTT(VCCIO), these are most important for overclocking, other voltages can help too.

Here is a table I put together from Intel&#8217;s product data sheet; it defines the processors maximum voltages and maximum amperage- These are OFFICIAL Min and Max Values:

(PICTURE SNIPPED)

VCC: Commonly Called Vcore, the voltage supplied to the processors inside the CPU. This voltage is a large part of Sandy Bridge overclocking. Now from personal testing, and weeks and weeks of headaches and hard ships, I have a few things I would like to share about this voltage.
I say maximum voltage for Vcc/Vcore is 1.50v for 24/7, 1.55-1.60v for extreme benchmarking, please stay below to 1.6, and don&#8217;t use any type of Load Line Calibration past 1.55v. Please not that high temperature teamed up with high voltage will kill your processor faster than anything else, it creates a perfect environment for processor degradation.

That's a lot more reliable to me than some random review site parroting Intel pr :). Guy knows his stuff from what I've seen in the past.
 
Please add to the list of important info:

LGA 1155 is Intel's new midrange socket. Midrange socket platforms come with limitations. If you attempt to use midrange equipment at extreme values, expect disappointment and failures. If you want a high-end socket platform with not so many limitations, wait for LGA 2011.
That will be the "real" Sandy Bridge and maybe will come at the "real" Sandy Bridge's pricing.. :D
 
That will be the "real" Sandy Bridge and maybe will come at the "real" Sandy Bridge's pricing.. :D

If it ends up being $350+, there's always Ivy Bridge soon after, entry level - midrange. I expect 5ghz on air without going over 1.40vcore.
 
Back
Top